Potential changes in the extent of suitable habitats for geladas (Theropithecus gelada) in the Anthropocene

Abstract

Background: Climate change coupled with other anthropogenic pressures may affect the extent of suitable habitat for species and thus their distributions. This is particularly true for species occupying high-altitude habitats such as the gelada (Theropithecus gelada) of the Ethiopian highlands. To explore the impact of climate change on species distributions, Species Distribution Modelling (SDM) has been extensively used. Here we model the current and future extent of sutibale habitat for geladas. Our modelling was based on 285 presence locations of geladas, covering their complete current distribution. We used different techniques to generate pseudoabsence datasets, MaxEnt model complexities, and cut-off thresholds to map the potential distribution of gelada under current and future climates (2050 and 2070). We assembled maps from these techniques to produce a final composite map. We also evaluated the change in the topographic features of gelada over the past 200 years by comparing the topography in current and historical settings.

Results: All model runs had high performances, AUC = 0.87-0.96. Under the current climate, the suitable habitat predicted with high certainty was 90,891 km², but it decreased remarkably under future climates, -36% by 2050 and - 52% by 2070. However, since the habitats of geladas already extend to mountaintop grasslands, no remarkable range shifts across elevation gradients were predicted under future climates.

Conclusions: Our findings indicated that climate change most likely results in a loss of suitable habitat for geladas, particularly south of the Rift Valley. Currently geladas are confined to higher altitudes and steep slopes compared to historical sightings, probably qualifying geladas as refugee species. The difference in topography is potentially associated with anthropogenic pressures that drove niche truncation to higher altitudes, undermining the climatic and topographic niche our models predicted. We recommend protecting the current habitats of geladas even when they are forecasted to become climatically unsuitable in the future, in particular for the population south of the Rift Valley.

Keywords: Climate change; Ethiopia; Habitat suitability modelling; High altitude primates; MaxEnt.

Fig. 1

Topographic map of Ethiopia indicating the relief and geographic positions of occurrence locations of geladas (Theropithecus gelada, red dots) after 2000. The broken lines encircle the assumed distribution ranges of the northern (T. g. gelada, green), the central (T. g. obscurus, yellow), and the southern (T. g. ssp. nov., orange) populations [20, 37]

Fig. 2

Gelada herd (Theropithecus gelada obscurus) in the Afro-alpine grassland in the highlands of cental Ethiopia (Guassa Community Conservation Area). Photos credit - Jeffrey T. Kerby

Fig. 3

Distribution and extent of suitable gelada habitat produced from 12 binary maps for current climate (current; 2 techniques to generate pseudo-absence points x 2 model complexity levels x 3 threshold values; see also Fig. S2), and from 36 binary maps for future scenarios (future 2050 and 2070); two techniques to generate pseudo-absence points x 2 model complexity levels x 3 threshold values x 3 emission scenarios). When grid cells in 30% or less of the binary maps (3 maps for current and 10 maps for future climates) predict suitability, we considered them unsuitable. When grid cells of > 30 − 60% maps (4–7 maps for the current climate and 11–21 maps for the future climate conditions) predicted suitability, we considered them uncertain in terms of suitability. When grid cells from > 60% binary maps (> 7 maps for the current and > 21 maps for the future) predict suitability, we considered them as suitable

Fig. 4

The area of suitable gelada habitats predicted under current and different future emission scenarios (2050 and 2070) produced using two levels of model complexity by setting a regularization multiplier values to 1 (Reg_1) and 8 (Reg_8) and by using three cut-off threshold values without and with bias file (WB): 10% (10 percentile omission rate), MTSS (maximum test sensitivity and specificity), and ETSS (equal test sensitivity and specificity). HG stands for Hadley Centre Global Environment Model version 2 (HadGEM2-ES). We used three emission scenarios (2.6, 4.5 and 8.5)

Fig. 5

Predicted change in habitat suitablities of geladas by 2050 (Curr_2050) and by 2070 (curr_2070). Green, pixels that are predicted to be suitable under both current and future climates; blue, pixels that are not currently predicted to be suitable but forecasted to be suitable in the future; yellow; currently suitable but not in the future; and grey, unsuitable both under current and future climates

Fig. 6

Changes in topographic features (elevation and slope) of gelada occurrence localities during the last 200 years. Box plots depict medians (horizontal lines), quartiles (box), ranges (whiskers), and means (red dots). The temporal variation in elevation of gelada occurrence locations is based on reported sighting from the 1800s, 1900s, and 2000s. For elevation change, we also added elevation of suitable habitat under the current climate (cur), and future climate conditions (2050 and 2070)